The Asymmetric Effect of Financial Development on Energy Consumption: NARDL Evidence from 173 Selected Countries

[Purpose] This study examines the asymmetric relationship between financial development (FD) and energy consumption (EC) across 173 economies from 1990 to 2019. It assesses whether positive and negative FD shocks affect EC differently and how these effects vary by income group, informing decisions that balance financial deepening and sustainable energy use. [Design/methodology/approach] A Nonlinear Autoregressive Distributed Lag (NARDL) framework is applied to a balanced panel of 56 high-income, 45 upper-middle-income, 50 lower-middle-income, and 22 low-income countries. FD is measured using the IMF Financial Development Index and decomposed into financial market development (FMD) and financial institution development (FID), with urbanisation and GDP per capita as controls. Positive and negative FD shocks are modeled via partial-sum processes, and pooled mean group estimators with Wald tests identify short- and long-run asymmetries. [Findings] Globally, positive shocks in overall FD and FMD reduce EC, whereas negative shocks in FD, FMD, and FID increase it. Heterogeneity is substantial: in high- and lower-middle-income economies, both positive and negative shocks to FD, FMD, and FID generally raise EC, while in upper-middle- and low-income economies, FMD tends to reduce EC under both shock types. Asymmetric effects are strongest for FD and FMD, with FID asymmetry concentrated in high-income countries. [Research limitations/implications] The estimates may not fully capture recent structural breaks in financial systems, technology, or energy policy. Notably, the dataset ends in 2019 and therefore excludes the substantial disruptions of the COVID-19 pandemic period. [Practical implications] Results suggest that steady financial deepening can curb EC, whereas contractions in FD may sharply increase it, especially in bank-dominated systems. Income-specific financial and energy policies are needed to manage these asymmetric effects. [Originality/value] This study is among the first to apply a global NARDL framework with the IMF FD index to model positive and negative FD, FMD, and FID shocks across 173 economies, linking asymmetric FD-EC dynamics to decision-relevant policy design.